What SAIDI and SAIFI mean

Utilities summarise a year of power interruptions with three indices, which they report to the U.S. Energy Information Administration on Form EIA-861.

• SAIDI, the System Average Interruption Duration Index: total customer-minutes of sustained interruption in the year divided by the number of customers served. A SAIDI of 120 means the average customer was without power for two hours over the year.
• SAIFI, the System Average Interruption Frequency Index: total customer interruptions divided by customers served. A SAIFI of 1.2 means the average customer lost power 1.2 times.
• CAIDI, the Customer Average Interruption Duration Index: SAIDI divided by SAIFI, the average length of one interruption.

SAIDI = sum of (customers interrupted × minutes) ÷ customers served

They are averages over a whole system. Most customers experience less than the average in a given year and a few experience far more.

Major event days

A hurricane, ice storm or wildfire shutoff can add more outage minutes in a week than a utility sees in several ordinary years. The IEEE 1366 standard defines a statistical threshold, based on the utility's own history, above which a day counts as a major event day. Utilities report SAIDI and SAIFI both with those days included and with them excluded. The first describes what customers lived through; the second is closer to the routine performance of the network.

Across the 893 utilities on this site with a 2024 filing, the median SAIDI was 184.5 minutes with major event days and 99.8 without.

Why the figures are only roughly comparable

• Method. Not every utility follows IEEE 1366. EIA's file records whether a respondent used the standard or another method, and this site shows which. Other methods may define a major event differently or not set one apart at all.
• What counts as sustained. The IEEE standard counts interruptions longer than five minutes. Some utilities use one minute, which raises their counts.
• Detection. Utilities with automated outage management record interruptions that others learn of only when a customer calls.
• Loss of supply. An outage caused by a failure on a transmission system the utility does not own may be included or reported separately.
• Territory and size. Long rural lines through forest are exposed to more weather per customer than underground urban networks. One storm can dominate the year of a utility with a few thousand customers.

For these reasons EIA publishes the figures "as reported", and so does this site. Pages show a utility's figure beside the median for its state and let the year-by-year table speak to whether a single year was unusual.

What the indices do not show

They say nothing about momentary flickers, voltage quality, where in a territory outages fall or how quickly a particular outage is restored. State utility commissions often publish more detailed reliability reports for the utilities they regulate.